Sync complete

models/RC/new/rel2id.json

@@ -1 +0,0 @@
-{"Product-Producer": 0, "Cause-Effect": 1, "Content-Container": 2, "Component-Whole": 3, "Other": 4, "Entity-Destination": 5, "Instrument-Agency": 6, "Entity-Origin": 7, "Message-Topic": 8, "Member-Collection": 9}

src/graph/GUI.py

@@ -20,8 +20,7 @@ sys.path.insert(0, default_path+'/../scripts')
 from src.scripts.functionsner import use_model, tag_sentence, json_to_txt, training_model, characterize_data, upsampling_data, usage_cuda, copy_data
 from src.scripts.functionsrc import  use_model_rc, training_model_rc, usage_cuda_rc 
 
-models = os.listdir(default_path+'/../../models')
-models.remove('RC')
+models_NER = os.listdir(default_path+'/../../models/NER')
 models_rc = os.listdir(default_path+'/../../models/RC')
 
 #-------------------------------------------Functions-----------------------------------------------
@@ -124,7 +123,7 @@ def Tagger_document_RC(Model, Input_file, Output_file, Cuda):
         
 #---------------------------------GUI-------------------------------------
 def execute_GUI():
-    global models
+    global models_NER
     with gr.Blocks(title='NER', css="#title {font-size: 150% } #sub {font-size: 120% } ") as demo:
         
         gr.Markdown("Named Entity Recognition(NER) and Relation Classification (RC) by GITA and Pratec Group S.A.S.",elem_id="title")
@@ -138,7 +137,7 @@ def execute_GUI():
                 with gr.Tab("Sentence"):
                     with gr.Row():
                         with gr.Column():
-                            b = gr.Radio(list(models), label='Model')
+                            b = gr.Radio(list(models_NER), label='Model')
                             inputs =[
                                  b,
                                  gr.Textbox(placeholder="Enter sentence here...", label='Sentence'), 
@@ -150,7 +149,7 @@ def execute_GUI():
                
                     
                     tagger_sen.click(Tagger_sentence, inputs=inputs, outputs=output)
-                    b.change(fn=lambda value: gr.update(choices=list(os.listdir('../../models')).remove('RC')), inputs=b, outputs=b)
+                    b.change(fn=lambda value: gr.update(choices=list(os.listdir('../../models/NER'))), inputs=b, outputs=b)
                     gr.Examples(
                     
                         examples=[
@@ -164,7 +163,7 @@ def execute_GUI():
                 with gr.Tab("Document"):
                     with gr.Row():
                         with gr.Column(): 
-                            c = gr.Radio(list(models), label='Model')
+                            c = gr.Radio(list(models_NER), label='Model')
                             inputs =[
                                  c,
                                  gr.File(label='Input data file'),
@@ -178,18 +177,18 @@ def execute_GUI():
                             gr.File(),
                             ]
                         
-                    models = os.listdir(default_path+'/../../models')
-                    models.remove('RC')
+                    models_NER = os.listdir(default_path+'/../../models/NER')
+                    
                     
                     tagger_json.click(Tagger_json, inputs=inputs, outputs=output)
-                    c.change(fn=lambda value: gr.update(choices=list(os.listdir('../../models')).remove('RC')), inputs=c, outputs=c)
+                    c.change(fn=lambda value: gr.update(choices=list(os.listdir('../../models/NER'))), inputs=c, outputs=c)
                     
              
             with gr.Tab("Trainer"):
                 with gr.Row():
                     with gr.Column():
                         train_input = inputs =[
-                             gr.Radio([True,False], label='Fast training', value=True),
+                             gr.Radio([True,False], label='Fast training', value=False),
                              gr.Textbox(placeholder="Enter model name here...", label='New model name'),
                              gr.Radio([True,False], label='Standard input', value=False),
                              gr.Textbox(placeholder="Enter path here...", label='Input data directory path'), 

src/scripts/functionsner.py

@@ -44,11 +44,11 @@ def str2bool(v):
         
 
 def copy_data(original_path):
-    data_folder  = default_path + '/../../data/train'
+    data_folder  = default_path + '/../../data/NER/train'
     copy_tree(original_path, data_folder)
     
 def characterize_data():
-    data_folder  = default_path + '/../../data/train'
+    data_folder  = default_path + '/../../data/NER/train'
     columns = {0: 'text', 1:'ner'}
     
     # init a corpus using column format, data folder and the names of the train, dev and test files
@@ -73,7 +73,7 @@ def characterize_data():
 
 def upsampling_data(entities_to_upsample, probability,  entities):
     print('-'*20,'upsampling','-'*20)
-    data_folder  = default_path + '/../../data/train'
+    data_folder  = default_path + '/../../data/NER/train'
     columns = {'text':0, 'ner':1}
     for m in ["SiS","LwTR","MR","SR", "MBT"]:
         upsampler = upsampling_ner(data_folder+'/train.txt', entities+['O'], columns)
@@ -107,8 +107,8 @@ def usage_cuda(cuda):
 def training_model(name, epochs=20):
     #FUNCION
     
-    data_folder  = default_path + '/../../data/train'
-    path_model = default_path + '/../../models/{}'.format(name)
+    data_folder  = default_path + '/../../data/NER/train'
+    path_model = default_path + '/../../models/NER/{}'.format(name)
     if (os.path.isdir(path_model)): print('WARNING, model already exists will be overwritten')
     columns = {0: 'text', 1:'ner'}
     # init a corpus using column format, data folder and the names of the train, dev and test files
@@ -192,7 +192,7 @@ def tag_sentence(sentence, name):
     
     
     #--------------Load the trained model-------------------------
-    path_model = default_path + '/../../models/{}'.format(name)
+    path_model = default_path + '/../../models/NER/{}'.format(name)
     global tagger_sentence
     
     if (not tagger_sentence):
@@ -239,7 +239,7 @@ def tag_sentence(sentence, name):
 def use_model(name, path_data, output_dir):
     
     #--------------Load the trained model-------------------------
-    path_model = default_path + '/../../models/{}'.format(name)
+    path_model = default_path + '/../../models/NER/{}'.format(name)
     
     if not (os.path.isdir(path_model)): 
         print('Model does not exists')
@@ -427,7 +427,7 @@ def json_to_txt(path_data_documents):
         id_in=groups_temp[0]
         
             
-        data_folder  = default_path + '/../../data/train'
+        data_folder  = default_path + '/../../data/NER/train'
         check_create(data_folder)
         count = 0
         with open(data_folder + '/{}.txt'.format(arch), mode='w', encoding='utf-8') as f: